UNSUPERVISED TEXTURE CLASSIFICATION USING VECTOR QUANTIZATION AND DETERMINISTIC RELAXATION NEURAL-NETWORK

This paper describes the use bf a neural network architecture for clas sifying textured images in an unsupervised manner using image-specific constraints, The texture features are extracted by using two-dimensio nal (2-D) Gabor filters arranged as a set of wavelet bases, The classi fication model comprises feature quantization, partition, and competit ion processes, The feature quantization process uses a vector Quantize r to quantize the features into codevectors, where the probability of grouping the vectors is modeled as Gibbs distribution, A set of label constraints for each pixel in the image are provided by the partition and competition processes. An energy function corresponding to the a p osteriori probability is derived from these processes, and a neural ne twork is used to represent this energy function, The state of the netw ork and the codevectors of tbe vector quantizer are iteratively adjust ed using a deterministic relaxation procedure until a stable state Is reached, The final equilibrium state of the vector quantizer gives a c lassification of the textured image. A cluster validity measure based on modified Hubert index is used to determine the optimal number of te xture classes in the image.